Skin color segmentation by histogram-based neural fuzzy network

Skin color image segmentation by a histogram-based self-constructing neural fuzzy inference network (SONFIN) is proposed in this paper. Each color pixel is represented by a hue-saturation (HS) space. To represent a block color by histogram as accurately as possible, a nonuniform quantization approach of HS space is considered. Histogram information of HS from images under different environments is used to train SONFIN to make the method as robust as possible. To verify performance of the proposed method, experiment on human-hand segmentation is performed. For comparison, other segmentation methods are applied to the same problem.